VASP running error with PGI compilers

[lv]

We compiled VASP 5.2 using
Red Hat Enterprise Linux Server release 5.8 (Tikanga) (64-bit)
VASP 5.2.12
Portland pgf90 10.3 64-bit
mpich2-1.4.1p1
fftw-3.1.2
GotoBLAS2-1.13
for parallelization on AMD Opteron(TM) Processor 6238 (Opteron Magny-Cours) 2.6 GHz, 4x12 cores. VASP crashed with the following output:

-----------------------------------------------------------------------------
| |
| ADVICE TO THIS USER RUNNING 'VASP/VAMP' (HEAR YOUR MASTER'S VOICE ...): |
| |
| You have a (more or less) 'small supercell' and for smaller cells |
| it is recommended to use the reciprocal-space projection scheme! |
| The real space optimization is not efficient for small cells and it |
| is also less accurate ... |
| Therefore set LREAL=.FALSE. in the INCAR file |
| |
-----------------------------------------------------------------------------

WARNING: for PREC=h ENMAX is automatically increase by 25 %
this was not the case for versions prior to vasp.4.4
WARNING: for PREC=h ENMAX is automatically increase by 25 %
this was not the case for versions prior to vasp.4.4
p6_9544: p4_error: interrupt SIGx: 4
p3_9173: p4_error: interrupt SIGx: 4
p7_9669: p4_error: interrupt SIGx: 4
p0_8837: p4_error: interrupt SIGx: 13
Killed by signal 2.
Killed by signal 2.
Killed by signal 2.
Killed by signal 2.
Killed by signal 2.
Killed by signal 2.
Killed by signal 2.
p0_8837: (21.355469) net_send: could not write to fd=4, errno = 32

========================================
Here attached with part of the makefile

FC=pgf90
# fortran linker
FCL=$(FC)
CPP_ = ./preprocess <$*.F | /usr/bin/cpp -P -C -traditional >$*$(SUFFIX)
CPP = $(CPP_) -DHOST=\"LinuxPgi\" \
-DNGXhalf -DCACHE_SIZE=2000 -DPGF90 -Davoidalloc \
-DRPROMU_DGEMV
FFLAGS = -I/home/opt/fftw3/include -Mfree -tp gh-64 -Kieee
OFLAG=-O2
BLAS= /home/opt/libgoto2_barcelonap-r1.13.a
LAPACK= ../vasp.5.lib/lapack_double.o
FC=mpif90 -I/home/opt/mpich2/include
CPP = $(CPP_) -DMPI -DHOST=\"LinuxPgi\" \
-DNGZhalf -DCACHE_SIZE=2000 -DPGF90 -Davoidalloc -DRPROMU_DGEMV
BLACS=/usr/local/BLACS_lam
SCA_= /usr/local/SCALAPACK_lam

SCA= $(SCA_)/scalapack_LINUX.a $(SCA_)/pblas_LINUX.a $(SCA_)/tools_LINUX.a \
$(BLACS)/LIB/blacsF77init_MPI-LINUX-0.a $(BLACS)/LIB/blacs_MPI-LINUX-0.a $(BLACS)/LIB/blacsF77init_MPI-LINUX-0.a

SCA=
LIB = -L../vasp.5.lib -ldmy \
../vasp.5.lib/linpack_double.o $(LAPACK) \
$(SCA) $(BLAS)

# FFT: only option fftmpi.o with fft3dlib of Juergen Furthmueller

FFT3D = fftmpi.o fftmpi_map.o fft3dfurth.o fft3dlib.o

However VASP is successfully compiled and running with ifort and mkl. Any other ideas I could try with PGI compiler?

Additionally, what would be optimal compiler and library for our new AMD machine? Do you have any suggestion about it?





<span class='smallblacktext'>[ Edited ]</span>